Skip to content
View thtjs0076's full-sized avatar

Block or report thtjs0076

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
thtjs0076/README.md

Hello, I'm Sosun Yim (Primula)

Data Science Student @ UC Berkeley
Data Scientist • NLP • Data Analytics • Backend Development


👋 About Me

I'm a Data Science student at UC Berkeley passionate about building practical AI systems that solve real-world problems.

My primary interests include:

  • 🤖 Machine Learning
  • 🧠 Natural Language Processing (NLP)
  • 📊 Data Analytics
  • 💻 Backend Development
  • 💁🏻 Data Scientist

I enjoy understanding data, building machine learning models, and making engineering decisions that balance accuracy, scalability, and real-world deployment.


🛠 Tech Stack

Languages

Machine Learning

  • Scikit-learn
  • Hugging Face Transformers
  • Pandas
  • NumPy
  • Matplotlib

Tools

  • Jupyter Notebook

🚀 Featured Projects

📧 Spam Classifier

Compared Logistic Regression and DistilBERT on 33K+ emails to evaluate the tradeoff between predictive performance and deployment efficiency.

Tech Stack

Python Scikit-learn Transformers Pandas

🔗 https://github.com/thtjs0076/spam-classifier


🌍 Multilingual TTS Router

Built a character-level language identification model using PyTorch and rebuilt it into a multilingual Text-to-Speech routing pipeline.

Tech Stack

Python PyTorch NLP

🔗 https://github.com/thtjs0076/multilingual-tts-router


📈 A/B Test Analysis

Performed an end-to-end A/B testing analysis to determine whether an observed lift was statistically significant and ready for production.

Tech Stack

Python Statistics Experimentation

🔗 https://github.com/thtjs0076/ab-test-analysis


🤖 Reinforcement Learning

Implemented Value Iteration and Q-Learning from scratch in a custom Gridworld environment.

Tech Stack

Python Reinforcement Learning

🔗 https://github.com/thtjs0076/rl-gridworld-qlearning


🌱 Currently Learning

  • Large Language Models (LLMs)
  • Retrieval-Augmented Generation (RAG)
  • MLOps
  • Production Machine Learning

📫 Connect

📧 Email

thtjs0076@gmail.com

💻 GitHub

https://github.com/thtjs0076

💼 LinkedIn

https://www.linkedin.com/in/sosun-yim-1a1533201

Popular repositories Loading

  1. spam-classifier spam-classifier Public

    Email spam classification: baseline to transformer

    Jupyter Notebook

  2. multilingual-tts-router multilingual-tts-router Public

    A language ID model from a class project, rebuilt into a real TTS routing pipeline for 5 languages

    Python

  3. ab-test-analysis ab-test-analysis Public

    An honest end-to-end look at one A/B test — is the lift real, and should we ship it?

    Jupyter Notebook

  4. rl-gridworld-qlearning rl-gridworld-qlearning Public

    Value iteration and Q-learning built from scratch on my own gridworld — and an honest look at where function approximation stops working

    Python

  5. thtjs0076 thtjs0076 Public